About Candle AI

Candle AI is a fast-growing legal-tech startup based in Austin, TX. Our AI-powered email assistant for lawyers integrates directly into Gmail & Outlook. Our mission is to revolutionize how high-volume professional services firms like law firms manage their client communications, save hours of billable time, and unlock additional revenue.

Backed by prominent investors like The Legal Tech Fund (TLTF) and with a growing pipeline of law firms & trusted early design partners, Candle AI is on track to become the status quo in legal email efficiency.

The Role

We’re seeking a Founding Backend Engineer with deep expertise in AI/ML systems and AWS serverless to architect and build the core infrastructure behind our intelligent email assistant. You’ll own critical systems—our RAG pipeline, LLM orchestration, Bedrock integrations, and multi-tenant serverless backend—and make architectural decisions that shape the product for years to come. As a founding team member, you’ll influence engineering culture, standards, and velocity from day one.

What You’ll Build

Serverless Backend Architecture (~50%)

• Architect serverless microservices powering core APIs and async workflows

• Design data models and noSQL DBs for multi-tenant SaaS with complex access patterns

• Build event-driven systems (SQS, RabbiqMQ)

• Implement OAuth 2.0 / OIDC integrations with 3P auth providers

• Own API Gateway configs, auth flows, rate limiting, and request shaping

• Optimize cold-starts, memory/CPU, concurrency, and overall cost efficiency

AI & Machine Learning Systems (~30%)

• Develop advanced LLM features using intelligent context orchestration and MCP-based workflows.

• Design and optimize RAG pipelines using AWS Bedrock Knowledge Bases and OpenSearch Serverless

• Implement embedding strategies and vector search for intelligent email context retrieval

• Create prompt orchestration and multi-step workflow frameworks with robust guardrails

• Develop evaluation/observability for AI quality, latency, and cost (success metrics, traces, red-team tests)

• Drive model selection, fine-tuning strategies, and cost/perf optimization

Infrastructure & DevOps (~20%)

• Define AWS CDK (TypeScript) IaC for all backend resources

• Configure VPC boundaries, security groups, IAM, and network policies where needed

• Implement monitoring, logging, and tracing (CloudWatch, X-Ray)

• Ship CI/CD with safe deploys and rollbacks

• Manage S3 data pipelines and ETL (incl. web crawlers where appropriate)

• Maintain and extend our SOC 2 Type II and Google CASA compliance posture, ensuring strong data privacy, security, and tenant isolation

Tech Stack You’ll Touch

AI/ML

• LLMs: Anthropic, OpenAI, Bedrock

• Vector Search: Experience with RAG Databases such as Pinecone and OpenSearch Serverless

• Frameworks/Patterns: Prompt engineering, tool-use/orchestration, evaluation harnesses

Backend Technologies

• Languages: Python 3.10+ (primary), TypeScript/Node.js (infra & tooling)

• TechStack: 2+ years of experience developing, deploying, and monitoring production services on AWS using multi-account strategies 

• Testing: pytest, integration tests; Jest for CDK unit tests

Tools & Practices

• Observability, distributed tracing, cost governance

What We’re Looking For

Required

• 5+ years backend engineering (Python&TypeScript-first), with production experience in serverless

• Hands-on with generative AI (Claude/OpenAI or equivalent) and RAG/vector search in prod

• Deep knowledge of AWS serverless (Lambda, DynamoDB, API Gateway, SQS/EventBridge)

• Strong AWS CDK or CloudFormation skills 

• Proven ability to design multi-tenant, scalable, and cost-efficient systems

• Implemented OAuth 2.0 / OIDC auth flows with providers like Google/Microsoft 

• Proficiency with AI coding tools (e.g., Cursor, Claude Code, GitHub Copilot, CodeX) where you leverage them to move faster without sacrificing code quality

• Excellent problem-solving, systems thinking, and clear communication

Highly Desired

• RAG & LLM Systems: Built production RAG pipelines using AWS Bedrock (or equivalent), vector DBs, and LLM eval frameworks.

• Data & Search: Strong in ETL, scraping, and semantic search/IR with deep vector and retrieval expertise.

• Cloud & Security: Advanced AWS (VPC, PrivateLink, Bedrock) with SOC 2 and CASA-compliant architectures.

• Model & OSS: Hands-on with model fine-tuning, OSS contributions, and thought leadership in AI/infra.

• Startup Impact: Proven founding-team experience shipping full-stack AI systems end-to-end.

Ideal Candidate Profile

• Tinker with the latest LLMs and ship pragmatic AI features

• Think in systems and trade-offs (latency, accuracy, cost, ops burden)

• Own outcomes from design → deploy → observe → iterate

• Thrive in ambiguity, ship quickly, and raise the bar on quality

• Collaborate tightly with the frontend to design ergonomic, durable APIs

Why This Role is Special

• Founding Impact: Shape architecture, processes, and team norms from day one

• AI-First Product: Build real-world RAG and LLM systems at scale

• Modern Stack: Serverless on AWS, Bedrock + frontier models with no legacy anchors

• High Autonomy: Own major systems end-to-end and see your work in production fast

Location & Work Style

• Location: Austin, TX (in-person)

• Team: Small, highly technical founding team

• Culture: Ownership, speed, and a bias toward shipping. The hiring process involves a trial period to evaluate mutual compatibility.

Compensation

Competitive salary + meaningful equity package commensurate with experience.

To Apply: Send your resume, GitHub, and a short note about your most interesting AI/ML system to careers@trycandle.ai

Optional prompts: (1) A system you built and what you measured, (2) Your RAG/vector search experience, (3) Why AI-powered email for legal excites you.

Candle AI is an equal opportunity employer. We celebrate diversity and are committed to building an inclusive team.